Announcement

**paulhelf** · Mar 25, 2009, 12:55 AM

Re: Configure Calcomp DrawingBoard III digitizing board as XInput device

Howdy again, folks:

Having not heard a peep from anyone in three weeks, I have continued on my own floundering adventure to progess on this issue, albeit slowly. I might have the problem almost licked, but I think I still would benefit from guidance with the next steps for successfully setting the DrawingBoard III to the active pointing device.

What I need help with will be clear at the end of this report.

----------------------------------------------

Here are the latest details, successes and failures:

1) I temporarily moved the digitizing board to an identical PC running Kubuntu 8.04 and confirmed that
the timeout issue is the same. I also had serendipitously over time installed a bunch of
libraries that are needed for compiling the calcomp driver from scratch in Kubuntu 8.04. These libraries are not currently on my Kubuntu 8.10 test machine...I started putting them there and got stuck on "xorg-server" (apt-get couldn't find the package).

2) I downloaded

http://xorg.freedesktop.org/archive/...p-1.1.0.tar.gz

and

http://xorg.freedesktop.org/archive/...p-1.1.2.tar.gz

and unpacked them separately. I was having trouble setting up xorg-server libraries needed to
compile version 1.1.2, but I was able to compile version 1.1.0 after transferring a bug fix (see
http://bugs.gentoo.org/204639) in version 1.1.2 to the 1.1.0 version of xf86Calcomp.c.

3) Not only am I not a LINUX expert, I don't know how to program in C, either (...I'm mostly an
end-user without funds to pay a professional programmer for help)....so I took my best guess that
the main issue truly was a timeout problem and I searched through xf86Calcomp.c for something
interesting to mangle.

On line 787 of xf86-input-calcomp/src/xf86Calcomp.c is

int retries = 5;

which I changed to

int retries = 100;

I successfully recompiled and copied

xf86-input-calcomp/src/.libs/calcomp_drv.so

to

/usr/lib/xorg/modules/input/calcomp_drv.so

4) I restarted X with new and wonderful results:

a) Now, /var/log/Xorg.0.log reports some good news:

---- snip-snip --------------
,
.
.
(II) Calcomp SetupProc called
(**) Option "Device" "/dev/ttyS0"
(II) CALCOMP driver: Serial device opened
(**) Option "MinX" "00000"
(**) Option "MaxX" "36000"
(**) Option "MinY" "00000"
(**) Option "MaxY" "24000"
(II) options read MaxX=36000, MaxY=24000
(II) Calcomp firmware ID : 70130C 09/18/91 16906-3
(II) Product ID : CalComp 3300
(II) Tablet size : 36000 x 7616
(II) Calcomp tablet queried OK
(**) Option "DeviceName" "TABLET"
(**) Calcomp device name is TABLET
(II) Calcomp base setup finished
.
.
.
(II) evaluating device (TABLET)
(II) XINPUT: Adding extended input device "TABLET" (type: TABLET)

-------------------------------

5) More news (good and bad):

"xinput --list TABLET" gives

"TABLET" id=3 [XExtensionPointer]
Num_buttons is 1
Num_axes is 3
Mode is Absolute
Motion_buffer is 0
Axis 0 :
Min_value is 0
Max_value is 36000
Resolution is 39400
Axis 1 :
Min_value is 0
Max_value is 7616
Resolution is 39400
Axis 2 :
Min_value is 0
Max_value is 32
Resolution is 32

----------------------------------
The good news is that xinput seems to at least see the device.
The bad news is that some of the attributes are wrong (for example,
on Axis 1, the Max_value should be 24000, not 7616 and the Num_buttons should
probably be 16 for the 16-button puck, not 1. I'm hopeful that these discrepancies can be easily corrected.

6) Bad news: "xinput set-pointer TABLET" fails:

X Error of failed request: BadDevice, invalid or uninitialized input device
Major opcode of failed request: 146 (XInputExtension)
Minor opcode of failed request: 12 (X_ChangePointerDevice)
Serial number of failed request: 12
Current serial number in output stream: 12

---------------------------------------------------------------

OK... I'll continue to hammer at this, but I'm on the XInput and xorg.conf learning curve
and I'd be grateful for any advice that might expedite my progress so that I don't spend another
three weeks just trying to get to the point where I can start using the hardware.

What I need help with is

1) Overcoming the "X Error" reported in item (6) so that I can successfully use "xinput set-pointer TABLET".

2) Configuring the correct hardware attributes (probably via some combination of xorg.conf,
xinput, and manual keys on the DrawingBoard III hardware configuration banks)

Thanks for your ear,
Paul H.

**paulhelf** · Mar 25, 2009, 10:55 AM

Re: Configure Calcomp DrawingBoard III digitizing board as XInput device

Howdy yet again,

It seems from the output of "xinput list TABLET" (see my previous reply to this thread) that the driver
is reading or manipulating some bits from the device OK, but not others.

I was able to "hardwire" the driver to report

a) that there are 16 buttons on the mouse (not 1 button)
b) that the AXIS 1 maximum dimension on the digitizing board is 24000 (not 7616)
c) that the board resolution is 1000 lpi (not 39400) in both AXIS 0 and AXIS 1

Unfortunately, these configuration data should not have to be hardwired into the driver. The driver really should be fixed by a proper programmer to determine and correctly report these from initilization data read from and/or written to the digitizing hardware itself and xorg.conf input options.

Additional Details:

DETAIL 1) NUMBER OF BUTTONS reported: Since the number of buttons is NOT a driver option
(like for the "mouse" driver), I didn't see an obvious way to set them [myself] except to continue to
mangle the calcomp driver. (If I knew how to write C programs, I'd add an xorg.conf option so the user
can set the number of buttons.)

On line 297 of xf86Calcomp.c, I changed
if (InitButtonClassDeviceStruct (dev, 1, map) == FALSE)

to

if (InitButtonClassDeviceStruct (dev, 16, map) == FALSE)

DETAIL 2) RESOLUTION reported: I changed lines 319 to 326 from

InitValuatorAxisStruct (dev, 0, priv->min_x, priv->max_x,
39400,
0 /* min_res */ ,
12000 /* max_res */ );
InitValuatorAxisStruct (dev, 1, priv->min_y, priv->max_y,
39400,
0 /* min_res */ ,
39400 /* max_res */ );

to

InitValuatorAxisStruct (dev, 0, priv->min_x, priv->max_x,
1000,
0 /* min_res */ ,
1000 /* max_res */ );
InitValuatorAxisStruct (dev, 1, priv->min_y, priv->max_y,
1000,
0 /* min_res */ ,
1000 /* max_res */ );

DETAIL 3) AXIS 1 BOARD LIMIT: I changed line 608 from

priv->max_y = (int)buffer[5] + ((int)buffer[4] << 7);
to

/* priv->max_y = (int)buffer[5] + ((int)buffer[4] << 7); */
priv->max_y = 24000;

DETAIL 4) After restarting X, "xinput list TABLET" now gives:

"TABLET" id=3 [XExtensionPointer]
Num_buttons is 16
Num_axes is 3
Mode is Absolute
Motion_buffer is 0
Axis 0 :
Min_value is 0
Max_value is 36000
Resolution is 1000
Axis 1 :
Min_value is 0
Max_value is 24000
Resolution is 1000
Axis 2 :
Min_value is 0
Max_value is 32
Resolution is 32

DETAIL 5) Further testing of the driver output with "xinput query-state TABLET"

a) When I place the 16-button puck at an arbitrary position on the digitizing board,
the driver correctly reports the position (in the "valuator" variables). For example

2 classes :
ButtonClass
button[1]=up
button[2]=up
button[3]=up
button[4]=up
button[5]=up
button[6]=up
button[7]=up
button[8]=up
button[9]=up
button[10]=up
button[11]=up
button[12]=up
button[13]=up
button[14]=up
button[15]=up
button[16]=up
ValuatorClass Mode=Absolute Proximity=In
valuator[0]=30473
valuator[1]=21813
valuator[2]=0

b) Holding button "0" down on the puck produces the following change in the output
of "xinput query-state TABLET":

button[1]=down

c) Holding any of buttons "4", "8", or "C" down on the puck produces the same change in the output
from "xinput query-state TABLET":

button[8]=down

d) Holding any of the other buttons down on the puck produces *no* change in the output from
"xinput query-state TABLET".

DETAIL 6: "xinput set-pointer TABLET" still produces

X Error of failed request: BadDevice, invalid or uninitialized input device
Major opcode of failed request: 146 (XInputExtension)
Minor opcode of failed request: 12 (X_ChangePointerDevice)
Serial number of failed request: 12
Current serial number in output stream: 12

Helpful suggestions would be welcome....

Thanks,
Paul H.

**paulhelf** · Mar 26, 2009, 12:02 AM

Re: Configure Calcomp DrawingBoard III digitizing board as XInput device

More success....

1) In spite of my only being an old FORTRAN user and not knowing C language, I managed to figure out the correct change in xf86Calcomp.c to fix the erroneously computed max_y values in Axis 1 (see "Detail 3" of my previous self-reply) -- that is, a fix that doesn't require hardwiring the max_y limit to 24000 for my hardware:

Instead of the "hardwire" change to xf86Calcomp.c from the aforementioned "Detail 3", change line 608 from

priv->max_y = (int)buffer[5] + ((int)buffer[4] << 7);
to

priv->max_y = (int)buffer[5] + ((int)buffer[4] << 7)
+ (((int)buffer[3]&0x03)<<14);

[shadow=red,left]This is a real bug fix. The continuation of line 608 is evidently missing from xf86Calcomp.c for aeons in the master source release (xf86-input-calcomp-master) from X.org, version 1.1.2, as well as version 1.1.0. I tested the change above and it correctly reports the y_max size of 24000 on my Calcomp 34360 DrawingBoard III without my having to hardwire the number in the driver.

2) Although "xinput set-pointer TABLET" still fails, I was able use

xsetpointer +c TABLET

to get the digitizing board to move the on-screen cursor and correctly respond when "button[1]" on the puck is pushed down. The on-screen dimensions were properly scaled to the digitizing board dimensions so that when I placed the puck in each active corner of the digitizing board, the cursor would move to the correct corner of the display.

3) I was able to use

xinput set-button-map "TABLET" 1 2 4 3 5 6 7 8 9 10 11 12 13 14 15 16

to get all of the cursor buttons to respond with xinput query-state TABLET

---------------------------------------------

This gets me nearly all the way to having a functional X-interface to my digitizing board. I haven't put it through its paces yet (I suspect that there may be some mouse attribute issues that will need to be solved -- for example, I noticed that I can't click and drag a window on the screen the way that I can with a regular three-button-mouse.). If I get this thing to work the way I want it to in Kubuntu 8.04, then the next step will be to see if I can get it working in Kubuntu 8.10.

If anyone out there is finding this thread useful, please reply.

I was sniffing around on the XFree86/X.org web sites and it appears that this calcomp driver isn't currently maintained (see http://cgit.freedesktop.org/xorg/dri...alcomp/commit/). I think the reasons that it isn't maintained are that it never worked right to begin with, the maintainers who inherited it didn't have the hardware to test it, and people like me who really could have used it were forced to find workarounds rather than struggling through doing it the right way.

If the only thing stopping maintainers from fixing this driver is that they don't have hardware on which to test it, I would be happy to test the driver fixes on my board. At present, I don't have a pressure pen (stylus) with my unit, but I do have 16-button pucks (wired and wireless).

If the case can be made that people need this driver, then perhaps we can persuade the Xorg maintainer guy to resurrect the driver and make it more user-friendly. If I were a C-programmer, I'd offer to maintain it myself...at least as long as I have a Calcomp DrawingBoard III on which to use it.

Cheers,
Paul H.

**paulhelf** · Mar 27, 2009, 03:36 PM

Re: Configure Calcomp DrawingBoard III digitizing board as XInput device

Howdy:

OK...I've moved back to Kubuntu 8.10 and I succeeded at compiling and installing my hacked version of the source code (xf86Calcomp.c) from

http://cgit.freedesktop.org/xorg/dri...-master.tar.gz

I'll attach my source file (xf86Calcomp.c.txt) below. I needed to install some libraries to get ./configure to work (notably, pkg-config, x11proto-input-dev, xserver-xorg-dev). As previously noted, "xinput set-pointer TABLET" still fails, but the following works fine:

xsetpointer +c TABLET
xinput set-button-map "TABLET" 1 2 4 3 5 6 7 8 9 10 11 12 13 14 15 16

I've done some simple testing with GIMP and ENGAUGE. I'm hoping it will be fine for GRASS and QGIS.

About the only thing I need to figure out now is how to get "button [1]" to click-and-drag the way the 3-button mouse does. It isn't a big deal, though, because the 3-button mouse is working at the same time.

----

I filed a bug report (Bug 20871) on http://bugs.freedesktop.org. Since this driver is currently unmaintained, I'm not sure we'll see any action soon. Hopefully, my thread here on the Kubuntu forum will be a useful alternative for any of you who want to use your Calcomp DrawingBoard II or III as an XInput device.

For anyone interested in cleaning this driver up, it really should have an input option (i.e. via xorg.conf or the corresponding HAL fdi file) to specify the number of buttons on the puck (if "Puck" is selected as an option for the "Cursor" input option). At some point I'll put together an fdi file to bypass the presumably obsolete xorg.conf way of doing things.

Cheers and good luck!
Paul Helfenstein

Attached Files

xf86Calcomp.c.txt (25.0 KB, 102 views)

Announcement

Configure Calcomp DrawingBoard III digitizing board as XInput device [solved]

Configure Calcomp DrawingBoard III digitizing board as XInput device [solved]

Comment

Comment

Comment

Comment